EP 185 - Bringing AI to the edge with analog computing
Industrial IoT Spotlight
00:00
The Cost of Latency in Machine Learning
Analyt computing uses embedded flash technology from, or earlier in the segment. Instead of storing just two levels, like one or zero, you can actually store multiple levels. And that allows us to have not just 10 megabytes of weights on chip, we actually have 80 million weights on chip. We don't need to run the entire neural network going shot-by-shot at a single time. So our latency is much faster than what you'll see in a digital system where you've got parts like ARVR or sorting systems.
Transcript
Play full episode
